BackgroundMistreatment of trainees remains a frequently reported phenomenon in medical education. One barrier to creating an educational culture of respect and professionalism may be a lack of alignment in the perceptions of mistreatment among different learners. Through the use of clinical vignettes, our aim was to assess the perceptions of trainees toward themes of potential mistreatment at different stages of training.MethodsBased on observations from external experts embedded in the clinical learning environment, six thematic areas of potential mistreatment were identified: verbal abuse, specialty-choice discrimination, non-educational tasks, withholding/denying learning opportunities, neglect and gender/racial insensitivity. Corresponding clinical vignettes were created and distributed to 1) medical students, 2) incoming interns, 3) residents/fellows. Perceptions of the appropriateness of the interactions depicted in the vignettes were measured on a 5-point Likert scale. Scores were categorized into neutral or appropriate (≤3) or inappropriate (i.e. mistreatment) (>3) and compared using chi-squared tests.ResultsFour hundred twenty seven trainees participated (182 students, 120 interns, 125 residents/fellows). Proportions of students perceiving mistreatment differed significantly from those of interns and residents/fellows in domains of verbal abuse, specialty discrimination and gender/racial insensitivity (p < 0.05). In scenarios comparing interns to residents/fellows, no significant differences were noted in perceptions of mistreatment in the domains of non-educational tasks, withholding learning and neglect.ConclusionsPerceptions of mistreatment differ at different developmental stages of medical training. After exposure to the clinical learning environment, perceptions of incoming interns did not differ from those of residents/fellows, implicating clinical rotations as a key period in indoctrinating students into the prevailing culture. More longitudinal studies are needed to confirm or better examine this phenomenon.Electronic supplementary materialThe online version of this article (doi:10.1186/s12909-016-0853-4) contains supplementary material, which is available to authorized users.